Stock price prediction via deep belief networks
University of New Brunswick
Artificial Intelligence (AI) techniques such as Neural Network (NN) have been widely used in the financial industry to predict stock prices to aid investment decisions. However, the traditional NN has been quickly surpassed by the new rapid developing Deep Belief Network (DBN) and its variants in terms of prediction accuracy in areas like image processing and speech recognition. This project aims to apply the DBN technique to stock price prediction and compare its performance with the traditional NN. In particular, we use the S&P500 index as a case study, and our numerical results show that DBN indeed performs better than the traditional NN. Hence, as a new generation of AI technique, DBN shows great promise in stock price prediction and forecasting.